Feature/mariadb galera monitoring

PR_DESCRIPTION.md

@@ -0,0 +1,97 @@
+**What this PR does / why we need it**:
+
+Feature: MariaDB Galera Cluster Status Monitoring
+
+This PR addresses issue #20755 by adding support for monitoring MariaDB Galera cluster status. It registers the standard Galera status variables as metrics in the Cloudpods monitoring system, enabling alerting and dashboards for cluster health. It also updates the Hostman service to support configuring Telegraf to collect these metrics from MySQL/MariaDB instances.
+
+### Implementation Details
+**1. `pkg/monitor/dbinit/measurements/mysql.go`**:
+Registered standard Galera status variables to `mysql` measurement definition:
+```go
+{
+"wsrep_cluster_size", "wsrep_cluster_size", monitor.METRIC_UNIT_COUNT,
+},
+{
+"wsrep_cluster_status", "wsrep_cluster_status", monitor.METRIC_UNIT_NULL,
+},
+{
+"wsrep_ready", "wsrep_ready", monitor.METRIC_UNIT_NULL,
+},
+{
+"wsrep_connected", "wsrep_connected", monitor.METRIC_UNIT_NULL,
+},
+```
+
+**2. `pkg/monitor/dbinit/measurements/metrics.csv`**:
+- Updated to include the new wsrep metrics keys.
+
+**3. `pkg/hostman/options/options.go`**:
+- Added `TelegrafMysqlServers` option to `SHostBaseOptions` to allow configuring MySQL connection strings for Telegraf.
+
+**4. `pkg/hostman/hostinfo/hostinfo.go`**:
+- Updated `OnCatalogChanged` to read `TelegrafMysqlServers` and configure the `mysql` input for Telegraf, specifically requesting `wsrep` string fields.
+
+**5. `pkg/hostman/system_service/telegraf.go`**:
+- Implemented generation of `[[inputs.mysql]]` configuration block in `telegraf.conf`.
+
+<!--
+- [ ] Smoke testing completed
+- [x] Unit test written
+-->
+
+**Does this PR need to be backport to the previous release branch?**:
+
+NONE
+
+<!--
+If no, just write "NONE".
+-->
+
+### Verification
+**1. `pkg/monitor/dbinit/measurements/mysql_test.go`**:
+Added unit test to verify metric presence:
+```go
+package measurements
+
+import (
+"testing"
+)
+
+func TestMysqlMetrics(t *testing.T) {
+metrics := map[string]bool{
+"wsrep_cluster_size":    false,
+"wsrep_cluster_status":  false,
+"wsrep_ready":           false,
+"wsrep_connected":       false,
+}
+
+for _, m := range mysql.Metrics {
+if _, ok := metrics[m.Name]; ok {
+metrics[m.Name] = true
+}
+}
+
+for name, found := range metrics {
+if !found {
+t.Errorf("metric %s not found in mysql definitions", name)
+}
+}
+}
+```
+
+**Automated Tests**:
+Run the verification test:
+```bash
+go test -v -run TestMysqlMetrics ./pkg/monitor/dbinit/measurements/
+```
+
+**Output**:
+```
+=== RUN   TestMysqlMetrics
+--- PASS: TestMysqlMetrics (0.00s)
+PASS
+ok  yunion.io/x/onecloud/pkg/monitor/dbinit/measurements0.002s
+```
+
+**2. Hostman Telegraf Configuration Verification**:
+Verified that `telegraf.conf` is correctly generated with `[[inputs.mysql]]` block when `TelegrafMysqlServers` is configured.

pkg/hostman/hostinfo/hostinfo.go

@@ -2571,6 +2571,17 @@ func (h *SHostInfo) OnCatalogChanged(catalog mcclient.KeystoneServiceCatalogV3)
 		conf["kafka"] = kafkaConf
 	}
 
+	if len(options.HostOptions.TelegrafMysqlServers) > 0 {
+		conf["mysql"] = map[string]interface{}{
+			"servers": options.HostOptions.TelegrafMysqlServers,
+			"string_fields": []string{
+				"wsrep_cluster_status",
+				"wsrep_ready",
+				"wsrep_connected",
+			},
+		}
+	}
+
 	urls, _ = s.GetServiceURLs("opentsdb", defaultEndpointType)
 	if len(urls) > 0 {
 		conf["opentsdb"] = map[string]interface{}{

pkg/hostman/options/options.go

@@ -67,6 +67,8 @@ type SHostBaseOptions struct {
 	TelegrafKafkaOutputSaslPassword  string `json:"telegraf_kafka_output_sasl_password" help:"telegraf kafka output sasl_password"`
 	TelegrafKafkaOutputSaslMechanism string `json:"telegraf_kafka_output_sasl_mechanism" help:"telegraf kafka output sasl_mechanism"`
 
+	TelegrafMysqlServers []string `json:"telegraf_mysql_servers" help:"telegraf mysql input servers"`
+
 	BackupTaskWorkerCount int `default:"3" help:"backup task worker count"`
 }
 

pkg/hostman/system_service/telegraf.go

@@ -309,6 +309,46 @@ func (s *STelegraf) GetConfig(kwargs map[string]interface{}) string {
 		conf += "\n"
 	}
 
+	if mysql, ok := kwargs["mysql"]; ok {
+		mysqlConf, _ := mysql.(map[string]interface{})
+		conf += "[[inputs.mysql]]\n"
+		if servers, ok := mysqlConf["servers"]; ok {
+			srvs, _ := servers.([]string)
+			serverStrs := make([]string, len(srvs))
+			for i, srv := range srvs {
+				serverStrs[i] = fmt.Sprintf("\"%s\"", srv)
+			}
+			conf += fmt.Sprintf("  servers = [%s]\n", strings.Join(serverStrs, ", "))
+		}
+		keys := []string{}
+		for k := range mysqlConf {
+			if k != "servers" {
+				keys = append(keys, k)
+			}
+		}
+		sort.Strings(keys)
+		for _, k := range keys {
+			v := mysqlConf[k]
+			switch val := v.(type) {
+			case string:
+				conf += fmt.Sprintf("  %s = \"%s\"\n", k, val)
+			case int, int32, int64:
+				conf += fmt.Sprintf("  %s = %d\n", k, val)
+			case float32, float64:
+				conf += fmt.Sprintf("  %s = %f\n", k, val)
+			case bool:
+				conf += fmt.Sprintf("  %s = %v\n", k, val)
+			case []string:
+				quoted := make([]string, len(val))
+				for i, s := range val {
+					quoted[i] = fmt.Sprintf("\"%s\"", s)
+				}
+				conf += fmt.Sprintf("  %s = [%s]\n", k, strings.Join(quoted, ", "))
+			}
+		}
+		conf += "\n"
+	}
+
 	if radontop, ok := kwargs[TELEGRAF_INPUT_RADEONTOP]; ok {
 		radontopMap, _ := radontop.(map[string]interface{})
 		devPaths := radontopMap[TELEGRAF_INPUT_RADEONTOP_DEV_PATHS].([]string)

pkg/monitor/dbinit/measurements/metrics.csv

@@ -506,6 +506,10 @@ Measurement,MeasurementNote,ResourceType,Database,Metric,MetricNote,MetricUnit
 "mysql","mysql","ext_mysql","telegraf","info_schema_table_rows","info_schema_table_rows","count"
 "mysql","mysql","ext_mysql","telegraf","info_schema_table_size_data_length","info_schema_table_size_data_length","count"
 "mysql","mysql","ext_mysql","telegraf","info_schema_table_size_index_length","info_schema_table_size_index_length","count"
+"mysql","mysql","ext_mysql","telegraf","wsrep_cluster_size","wsrep_cluster_size","count"
+"mysql","mysql","ext_mysql","telegraf","wsrep_cluster_status","wsrep_cluster_status","NULL"
+"mysql","mysql","ext_mysql","telegraf","wsrep_ready","wsrep_ready","NULL"
+"mysql","mysql","ext_mysql","telegraf","wsrep_connected","wsrep_connected","NULL"
 "netstat","netstat","host","telegraf","tcp_established","","count"
 "netstat","netstat","host","telegraf","tcp_syn_sent","","count"
 "netstat","netstat","host","telegraf","tcp_syn_recv","","count"

pkg/monitor/dbinit/measurements/mysql.go

@@ -54,5 +54,17 @@ var mysql = SMeasurement{
 		{
 			"info_schema_table_size_index_length", "info_schema_table_size_index_length", monitor.METRIC_UNIT_COUNT,
 		},
+		{
+			"wsrep_cluster_size", "wsrep_cluster_size", monitor.METRIC_UNIT_COUNT,
+		},
+		{
+			"wsrep_cluster_status", "wsrep_cluster_status", monitor.METRIC_UNIT_NULL,
+		},
+		{
+			"wsrep_ready", "wsrep_ready", monitor.METRIC_UNIT_NULL,
+		},
+		{
+			"wsrep_connected", "wsrep_connected", monitor.METRIC_UNIT_NULL,
+		},
 	},
 }

pkg/monitor/dbinit/measurements/mysql_test.go

@@ -0,0 +1,26 @@
+package measurements
+
+import (
+	"testing"
+)
+
+func TestMysqlMetrics(t *testing.T) {
+	metrics := map[string]bool{
+		"wsrep_cluster_size":    false,
+		"wsrep_cluster_status":  false,
+		"wsrep_ready":           false,
+		"wsrep_connected":       false,
+	}
+
+	for _, m := range mysql.Metrics {
+		if _, ok := metrics[m.Name]; ok {
+			metrics[m.Name] = true
+		}
+	}
+
+	for name, found := range metrics {
+		if !found {
+			t.Errorf("metric %s not found in mysql definitions", name)
+		}
+	}
+}